SDK for stock analysis and strategy backtest.
Project description
pqsdk
介绍
品宽量化交易平台提供的本地化SDK，支持使用熟悉的工具进行量化程序开发与回测
安装方法
# 首次安装, 在项目的根目录下添加requirements.txt文件，增加内容如下：
Cython==3.0.11
plotly-express==0.4.1
msgpack>=0.4.7
pandas==1.5.2
requests==2.31.0
six==1.16.0
thriftpy2==0.5.2
colorlog==4.8.0
QuantStats==0.0.62
ipython==8.12.2
numpy==1.24.4
tqdm==4.65.0
pqsdk
# 在命令行执行：
pip install -r requirements.txt
# 如果安装失败，请指定PIP官方镜像源
pip install -r requirements.txt -i https://pypi.org/simple
# 安装指定的版本： pip install pqsdk==<version>
pip install pqsdk==0.0.42
# 升级版本
pip install -U pqsdk
# 如果安装失败，请指定官方镜像源
pip install -U pqsdk -i https://pypi.org/simple
使用方法
from pqsdk.api import *
"""账号认证, 以品宽量化交易平台www.pinkquant.com的用户名/密码登录"""
username = 'vivian'
password = 'mypassword'
auth(username, password)
"""使用 token 认证账号"""
auth_by_token(token="")
# 获取因子数据
factor_list = ['float_share', 'pe', 'pe_ttm', 'ma_5', 'ema_5', 'dv_ratio']
df = get_factor(stock_pool=['000300.SH'], trade_date='2023-03-29', factor_list=factor_list)
print(df)
open high ... volume amount
sec_code trade_date ...
000001.SZ 2023-03-29 12.73 12.74 ... 596064.33 750687.551
000002.SZ 2023-03-29 15.41 15.62 ... 654848.26 1012734.687
000063.SZ 2023-03-29 34.435106 34.553575 ... 782218.36 2685933.703
000069.SZ 2023-03-29 4.7 4.75 ... 324836.19 152639.669
000100.SZ 2023-03-29 3.989899 4.044431 ... 1604844.15 707678.236
... ... ... ... ... ...
688363.SH 2023-03-29 111.999999 113.159999 ... 30500.33 341085.693
688396.SH 2023-03-29 59.307833 61.559399 ... 75640.44 458018.471
688561.SH 2023-03-29 72.570005 72.570005 ... 87323.24 605371.703
688599.SH 2023-03-29 51.750011 52.490011 ... 149553.3 768914.461
688981.SH 2023-03-29 49.189988 50.319988 ... 708926.22 3496451.834
[300 rows x 6 columns]
open high ... volume amount
sec_code datetime ...
000001.SZ 2023-03-29 14:56:00 12.529999 12.54 ... 3274.0 4103280.0
2023-03-29 14:57:00 12.54 12.54 ... 5943.0 7450969.0
2023-03-29 14:58:00 12.54 12.54 ... 122.0 152976.0
2023-03-29 14:59:00 12.54 12.54 ... 0.0 0.0
2023-03-29 15:00:00 12.54 12.54 ... 9497.0 11900655.0
[5 rows x 6 columns]
从命令行进行回测
从Terminal命令行启动的格式：
pqsdk backtest -f <策略程序文件> -p <key=value> --plot --save_results --save_orders
系统预设了如下关于-p <key=value>的参数：
| 参数 | 默认值 | 说明 |
| --- | --- | --- |
| cash | 1000000 | 回测初始资金 |
| start_date | 2024-03-01 | 回测开始日期 |
| end_date | 2024-03-20 | 回测结束日期 |
| benchmark | 000300.SH | 回测基准 |
| stock_pool | 000300.SH,000905.SH | 股票池，即指数代码，多个指数代码用逗号分隔 |
| unit | 1d | 基准单位，支持1d，1m |
| adjust_period | 5 | 调仓周期 |
| hold_maxsize | 10 | 最大持仓股数 |
-plot参数为布尔值，是否要生成累计收益率的Tearsheet，不指定时为False。
案例：
pqsdk backtest -f tests/buy_and_hold.py -p cash=1000000 -p start_date=2024-03-01 -p end_date=2024-03-20 -p benchmark=000300.SH -p stock_pool=000300.SH,000905.SH -p unit=1d -p adjust_period=5 -p hold_maxsize=10 -plot
pqsdk backtest -f tests/buy_and_hold.py -p start_date=2024-03-01 -p end_date=2024-03-20 -p benchmark=000300.SH -p stock_pool=000300.SH,000905.SH
pqsdk backtest -f stop_loss_strategy.py -p cash=1000000 -p start_date=2024-03-01 -p end_date=2024-03-17 -p benchmark=000300.SH -p stock_pool=000300.SH,000905.SH -p unit=1d -p adjust_period=5 -p hold_maxsize=10 -plot
-save_metrics 是否保存结果到csv文件
-save_orders 是否保存委托到csv文件
从代码进行回测
代码案例：
from pqsdk.backtest import execute, tearsheet, save_metrics, save_orders, save_inout_cash
from pqsdk import log
from pqsdk.api import *
# 回测参数
params = {
"cash": 1000000,
"start_date": "2024-08-20",
"end_date": "2024-09-03",
"benchmark": "000300.SH",
"stock_pool": "000300.SH".split(","),
"unit": '1d',
"adjust_period": 5,
"hold_maxsize": 10,
}
#
strategy_file = "buy_and_hold.py"
# 执行回测
results = execute(parameters=params, strategy_file=strategy_file)
# print(results)
tearsheet(results)
save_metrics(results)
save_orders(results)
save_inout_cash(results)
回测结果
回测结果是一个字典，包含以下内容的案例：
results = {
'benchmark': '000300.SH',
'analysis': {'time_return': {...}},
'plot_data': {},
'orders':
order_id sec_code volume price is_buy avg_cost comm add_time trade_date
0 1 000533.SZ 62800 4.769999 True 4.769999 25.579494 2024-01-03 15:00:00 2024-01-03
1 2 000533.SZ 62800 4.599999 False 4.769999 313.726004 2024-01-08 15:00:00 2024-01-08
2 3 000417.SZ 59600 4.973573 True 4.973573 25.364394 2024-01-10 15:00:00 2024-01-10
3 4 000417.SZ 59600 4.769180 False 4.973573 308.770614 2024-01-17 15:00:00 2024-01-17,
'inout_cash':
datetime cash
0 2024-01-01 1000000}
其中:
- results['analysis']['time_return']['strat_return'] 为策略每日收益率。
- results['analysis']['time_return']['bchmk_return'] Benchmark的收益率，与策略收益率strat_return进行对比.
'strat_return'为pd.Series，样例:
2024-01-02 0.000000e+00
2024-01-03 0.000000e+00
2024-01-04 0.000000e+00
2024-01-05 -5.651995e-03
2024-01-08 -5.052566e-03
2024-01-09 0.000000e+00
2024-01-10 2.220446e-16
2024-01-11 4.104443e-03
2024-01-12 -5.839526e-03
2024-01-15 7.635953e-03
2024-01-16 -4.663432e-03
2024-01-17 -1.347019e-02
2024-01-18 0.000000e+00
2024-01-19 0.000000e+00
2024-01-22 0.000000e+00
2024-01-23 0.000000e+00
2024-01-24 0.000000e+00
'bchmk_return'为pd.Series，样例:
2024-01-02 -0.013045
2024-01-03 -0.002379
2024-01-04 -0.009249
2024-01-05 -0.005360
2024-01-08 -0.012933
2024-01-09 0.001961
2024-01-10 -0.004668
2024-01-11 0.005658
2024-01-12 -0.003491
2024-01-15 -0.000989
2024-01-16 0.006083
2024-01-17 -0.021750
2024-01-18 0.014136
2024-01-19 -0.001512
2024-01-22 -0.015559
2024-01-23 0.004048
2024-01-24 0.013978
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
pqsdk-0.0.67.tar.gz
(65.7 kB
view details)
File details
Details for the file pqsdk-0.0.67.tar.gz.
File metadata
- Download URL: pqsdk-0.0.67.tar.gz
- Upload date:
- Size: 65.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.0.0 CPython/3.8.19
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
bb9d32f15c9ad03f7aad4033c80c3d88cb67b9a90cf72226c628734da31b875a
|
|
| MD5 |
b9543a7ad8cc8aa8b32308c88df768e6
|
|
| BLAKE2b-256 |
c53accd091df1eed15b2c319630ce2ccc2147dde551759fcd40fbaa5896849f3
|
